0% found this document useful (0 votes)
12 views

Exercise Class Week3 2015

The document discusses three statistical mechanics questions: calculating the volume of a D-dimensional sphere, analyzing the behavior of an ideal gas, and comparing systems with bounded and unbounded energy spectra. Q6 asks to derive an expression for volume of a D-dimensional sphere. Q7 discusses that derivatives of the entropy equation for an ideal gas yield simple ideal gas behavior. Q8 contrasts heat capacity for systems with bounded versus unbounded single-particle energy levels.
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
12 views

Exercise Class Week3 2015

The document discusses three statistical mechanics questions: calculating the volume of a D-dimensional sphere, analyzing the behavior of an ideal gas, and comparing systems with bounded and unbounded energy spectra. Q6 asks to derive an expression for volume of a D-dimensional sphere. Q7 discusses that derivatives of the entropy equation for an ideal gas yield simple ideal gas behavior. Q8 contrasts heat capacity for systems with bounded versus unbounded single-particle energy levels.
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

PHYS4031 STATISTICAL MECHANICS

SAMPLE QUESTIONS FOR DISCUSSION in WEEK 3 EXERCISE CLASSES


(21-23 September 2015)
You may want to think about them before attending exercise class.
SQ6 - Volume of a D-dimensional sphere of radius R
SQ7 - Inspecting S(V ) for classical ideal gas behavior
SQ8 - Two-level systems (bounded single-particle spectrum) differ qualitatively from collection of oscillators (un-
bounded single-particle spectrum)

SQ6 Mathematical Skills V – Volume of a D-dimensional “sphere”.


You know the expressions πR2 and 43 πR3 . How about higher dimensions? We need the answer in studying
classical ideal gas within the microcanonical ensemble.
The volume of a hyper-sphere of radius R in D dimensions is formally given by
∫ ∫
VD (R) = · · · dx1 · · · dxD . (1)
0≤x21 +···x2D ≤R2

(Why is it?) We want to carry out this integral. This problem makes use of the Gaussian integral and the
Gamma function, as discussed in Problem Set 1.
To proceed, by dimensional analysis, the surface area of a hyper-sphere of radius r can be written as:

AD (r) = SD rD−1 (2)

where SD is a constant to be determined. If we can find SD , then VD (R) follows immediately since
∫ R
SD D
VD (R) = SD rD−1 dr = R . (3)
0 D

Thus the whole problem becomes that of determining SD . Eq. (3) amounts to adding up shells as in an onion.
To get SD , we consider the following integral:
∫ ∞ ∫ ∞ ∫ ∞
dxD e−(x1 +x2 +···+xD )
2 2 2
ID ≡ dx1 dx2 · · · (4)
−∞ −∞ −∞

where the integrals are over all space in D dimensions.

(a) Noting that ID is a product of Gaussian integrals, show that

ID = π D/2 . (5)

(b) The integrand in ID is “spherically symmetrical” (just like the spherically symmetrical potentials you
encountered in atomic problems or in central force problems) and this allows us to treat it in a different
way. Treating the integral over all D-dimensional space by adding up “spherical shells” of different radii,
show that
SD D
ID = Γ( ) (6)
2 2
where Γ is the Gamma Function (see Problem Set 1 about Gamma functions).
(c) Hence, show that the volume of a D-dimensional hypersphere is given by

π D/2
VD (R) = RD . (7)
Γ( D
2 + 1)

Remarks: This formula is needed for studying classical ideal gas within microcanonical ensemble (if we
do not want to invoke any particle-in-a-big-box QM). It can also be used to study a collection of classical
oscillators. To get a feeling of the general formula, write down the explicit results for D = 1, 2, 3, 4. I
think one can also do a mathematical induction proof. I hope that you are happy as you have just learnt
something that goes beyond high school mathematics!

1
SQ7 Something that looks complicated could be simple - always inspects your answer – Ideal gas
behavior.
After a laborious calculation, one finds that the entropy of a 3D classical ideal gas is given by
[( )N ( )3N/2 ]
V 4πmE 5
S(E, V, N ) = k ln + Nk (8)
N 3h2 N 2
This is the Sackur-Tetrode Equation. This is exact, but looks complicated and scary. Some students then
become reluctant to take derivatives of S.
TA: Point out that things can be easy. If we want to find the pressure p or p/T , all we need is a derivative
with respect to V . Inspect Eq. (8). If we extract the V -dependence, what we have is simply:
S(V ) ∼ k ln(B V N )
∼ N k ln(B V ) (9)
where B is a big trunk of things, but irrelevant to our goal of taking a derivative w.r.t. V .
Find p/T and see how pV is related to the other parameters. Inspect the result and Eq. (8) and Eq. (9) again.
There are by-products. (a) Whenever you see the form of S has V -dependence given by Eq. (9), then we have
the ideal gas behavior. (b) This is informative – it implies when we have non-ideal gas behavior, i.e., when
there is some interaction between between the particles (atoms/molecules) in a gas, then we will not get the
ln V N form in S. If you trace the origin of this V N , it comes from the fact that the Hamiltonian of an ideal gas
does not depend on the positions (because of no interaction terms)! Everything comes out nice and consistent.
Students: For µ/T , there is no easy path but to work it out carefully, as there are three places where N appears
in S. Try it!
SQ8 Very important – Two-level Systems and features due to bounded single-particle energy spec-
trum.
In Problem 1.5, you worked out how M units of energy quanta (h̄ω) could be distributed among N distin-
guishable oscillators and each of them can take on 0, 1, 2, 3, . . . units of energy. The important point is that
the single-particle energy spectrum is unbounded, i.e., no ceiling on how much energy an oscillator can take.
You then worked out the heat capacity C(T ).
There is another class of problems in which the single-particle energy spectrum is bounded, i.e., there is a
ceiling. The key point is to know that they exhibit qualitatively different behavior.
Consider a system consisting of N independent (i.e., not interacting) particles with N ≫ 1. Each particle can
be in one of two energy levels ϵ1 = 0 and ϵ2 = ϵ > 0. [Many physical problems belong to this class. Consider
magnetic moments (atoms may have magnetic moments) sitting at their equilibrium positions in a solid. When
a magnetic field is applied, Zeeman splitting leads to a finite number of levels for each moment. For J = 1/2,
we have two levels for each atom. Another example is the Schottky defect problem in solids where ϵ1 = 0 refers
to an atom sitting in its crystalline (right) position and ϵ2 = ϵ is the excitation of an atom to a wrong position.]
(a) Given ϵ1 and ϵ2 , a macrostate can be described by E and N . To satisfy the given E and N , we need to
put N1 particles in level ϵ1 and N2 particles in level ϵ2 . Find N1 and N2 , in terms of E, N , and ϵ.
(b) Counting comes in. We need to divide N into two groups N1 and N2 . Find the number of microstates
W (E, N ). Hence obtain an expression for the entropy S(E, N ). Simplify the expression using the Stirling’s
formula, assuming N ,N1 ,N2 ≫ 1.
(c) Derive an expression for the temperature T , in terms of N1 and N2 . Hence, obtain N2 /N1 as a function
of temperature and sketch the result. Next, turn the result into E(T ), and obtain the heat capacity by
dE/dT . Sketch C(T ). The form of C(T ) is called the Schottky heat capacity. Key Point: Contrast
Schottky C(T ) (bounded spectrum) with that of a collection of oscillators (unbounded spectrum), as
obtained in Problem 1.5. These are the TWO classes of C(T ) behavior in physics – they are fundamental
different.
Important Remarks (TA: don’t need to do these in exercise classes):
– With bounded spectrum, C(T ) goes up and down and shows a peak.
– There is also an unusual behavior in the (derived) “temperature” as a function of N1 /N2 . What happens
if we have N2 > N1 ?
– Students may want to work out the more general answer in terms of ϵ1 and ϵ2 .
– You may try 3-level, 4-level,... systems. They are just further examples of bounded single-particle levels.

You might also like